package Perl::Critic::Policy::ValuesAndExpressions::ProhibitCommaSeparatedStatements; use 5.010001; use strict; use warnings; use Readonly; use Perl::Critic::Utils qw{ :booleans :characters :severities :classification hashify }; use Perl::Critic::Utils::PPI qw{ is_ppi_statement_subclass }; use parent 'Perl::Critic::Policy'; our $VERSION = '1.152'; #----------------------------------------------------------------------------- Readonly::Scalar my $DESC => q{Comma used to separate statements}; Readonly::Scalar my $EXPL => [ 68, 71 ]; #----------------------------------------------------------------------------- sub supported_parameters { return ( { name => 'allow_last_statement_to_be_comma_separated_in_map_and_grep', description => 'Allow map and grep blocks to return lists.', default_string => $FALSE, behavior => 'boolean', }, ); } sub default_severity { return $SEVERITY_HIGH } sub default_themes { return qw( core bugs pbp certrule ) } sub applies_to { return 'PPI::Statement' } #----------------------------------------------------------------------------- sub violates { my ( $self, $elem, undef ) = @_; # Grrr... PPI instantiates non-leaf nodes in its class hierarchy... return if is_ppi_statement_subclass($elem); # Now, if PPI hasn't introduced any new PPI::Statement subclasses, we've # got an element who's class really is PPI::Statement. return if _is_parent_a_constructor_or_list($elem); return if _is_parent_a_for_loop($elem); if ( $self->{_allow_last_statement_to_be_comma_separated_in_map_and_grep} ) { return if not _is_direct_part_of_map_or_grep_block($elem); } foreach my $child ( $elem->schildren() ) { if ( not $self->{_allow_last_statement_to_be_comma_separated_in_map_and_grep} and not _is_last_statement_in_a_block($child) ) { if ( $child->isa('PPI::Token::Word') ) { return if _succeeding_commas_are_list_element_separators($child); } elsif ( $child->isa('PPI::Token::Operator') ) { if ( $child->content() eq $COMMA ) { return $self->violation($DESC, $EXPL, $elem); } } } } return; } sub _is_parent_a_constructor_or_list { my ($elem) = @_; my $parent = $elem->parent(); return if not $parent; return ( $parent->isa('PPI::Structure::Constructor') or $parent->isa('PPI::Structure::List') ); } sub _is_parent_a_for_loop { my ($elem) = @_; my $parent = $elem->parent(); return if not $parent; return if not $parent->isa('PPI::Structure::For'); return 1 == scalar $parent->schildren(); # Multiple means C-style loop. } sub _is_direct_part_of_map_or_grep_block { my ($elem) = @_; my $parent = $elem->parent(); return if not $parent; return if not $parent->isa('PPI::Structure::Block'); my $block_prior_sibling = $parent->sprevious_sibling(); return if not $block_prior_sibling; return if not $block_prior_sibling->isa('PPI::Token::Word'); state $is_map_or_grep = { hashify( qw( map grep ) ) }; return $is_map_or_grep->{$block_prior_sibling}; } sub _is_last_statement_in_a_block { my ($elem) = @_; my $parent = $elem->parent(); return if not $parent; return if not $parent->isa('PPI::Structure::Block'); my $next_sibling = $elem->snext_sibling(); return if not $next_sibling; return 1; } sub _succeeding_commas_are_list_element_separators { my ($elem) = @_; if ( is_perl_builtin_with_zero_and_or_one_arguments($elem) and not is_perl_builtin_with_multiple_arguments($elem) ) { return; } my $sibling = $elem->snext_sibling(); return 1 if not $sibling; # There won't be any succeeding commas. return not $sibling->isa('PPI::Structure::List'); } 1; __END__ #----------------------------------------------------------------------------- =pod =head1 NAME Perl::Critic::Policy::ValuesAndExpressions::ProhibitCommaSeparatedStatements - Don't use the comma operator as a statement separator. =head1 AFFILIATION This Policy is part of the core L distribution. =head1 DESCRIPTION Perl's comma statement separator has really low precedence, which leads to code that looks like it's using the comma list element separator not actually doing so. Conway suggests that the statement separator not be used in order to prevent this situation. The confusion that the statement separator causes is primarily due to the assignment operators having higher precedence. For example, trying to combine two arrays into another like this won't work: @x = @y, @z; because it is equivalent to @x = @y; @z; Conversely, there are the built-in functions, like C, that normally force the rest of the statement into list context, but don't when called like a subroutine. This is not likely to produce what is intended: print join q{, }, 2, 3, 5, 7, ": the single-digit primes.\n"; The obvious fix is to add parentheses. Placing them like print join( q{, }, 2, 3, 5, 7 ), ": the single-digit primes.\n"; will work, but print ( join q{, }, 2, 3, 5, 7 ), ": the single-digit primes.\n"; will not, because it is equivalent to print( join q{, }, 2, 3, 5, 7 ); ": the single-digit primes.\n"; =head1 CONFIGURATION This policy can be configured to allow the last statement in a C or C block to be comma separated. This is done via the C option like so: [ValuesAndExpressions::ProhibitCommaSeparatedStatements] allow_last_statement_to_be_comma_separated_in_map_and_grep = 1 With this option off (the default), the following code violates this policy. %hash = map {$_, 1} @list; With this option on, this statement is allowed. Even if this option is off, using a fat comma C<< => >> works, but that forces stringification on the first value, which may not be what you want. =head1 BUGS Needs to check for C. =head1 AUTHOR Elliot Shank C<< >> =head1 COPYRIGHT Copyright (c) 2007-2023 Elliot Shank. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. The full text of this license can be found in the LICENSE file included with this module. =cut # Local Variables: # mode: cperl # cperl-indent-level: 4 # fill-column: 78 # indent-tabs-mode: nil # c-indentation-style: bsd # End: # ex: set ts=8 sts=4 sw=4 tw=78 ft=perl expandtab shiftround :